bestofgithub.dev

← Back to the directory

OpenPipe/ART

View on GitHub

Agent Reinforcement Trainer: train multi-step agents for real-world tasks using GRPO. Give your agents on-the-job training. Reinforcement learning for Qwen3.6, GPT-OSS, Llama, and more!

OpenPipe/ART is #5,573 on the bestofgithub Elo ladder, in Silver II, with a rating of 1,493 (−10 this week). It has 10,857 stars, 9 gained in the last week.

Silver II

1,493 Elo

  • #5,573 on the ladder
  • −6 today · −10 this week
  • Peak 1,509
8 Elo to Silver I

Stars

10,857

  • +1 today
  • +9 this week
  • – this month

Repository

Forks
1k
Language
Python
Last push
Category
AI & Agents

Achievements

bestofgithub badgeMaintainer? Show your league in your README

Card: your league's emblem, from Copper to Legendary

[![bestofgithub](https://bestofgithub.dev/badge/OpenPipe/ART.svg?style=card)](https://bestofgithub.dev/repo/OpenPipe/ART)

Flat: fits beside your other badges

[![bestofgithub](https://bestofgithub.dev/badge/OpenPipe/ART.svg)](https://bestofgithub.dev/repo/OpenPipe/ART)

Install

git
git clone https://github.com/OpenPipe/ART.git

Similar repositories

oumi-ai/oumi

Open on GitHubPython

Easily fine-tune (SFT/RL), evaluate, and deploy Qwen, Gemma, or any open agentic LLM/VLM!

  • 1,430+3 Elo, change today
  • 9.4k stars
  • 798 forks
git
git clone https://github.com/oumi-ai/oumi.git
  • agentic-ai
  • dpo
  • fine-tuning
  • gemma

HKUDS/AutoAgent

Open on GitHubPython

"AutoAgent: Fully-Automated and Zero-Code LLM Agent Framework"

  • 1,469+2 Elo, change today
  • 9.8k stars
  • 1.3k forks
git
git clone https://github.com/HKUDS/AutoAgent.git
  • agent
  • llms

modelscope/ms-swift

Open on GitHubPython

Use PEFT or Full-parameter to CPT/SFT/DPO/GRPO 600+ LLMs (Qwen3.8, DeepSeek-V4, GLM-5.1, InternLM3, Llama4, ...) and 300+ MLLMs (Qwen3-VL, Qwen3-Omni, InternVL3.5, Ovis2.5, GLM5.3, Gemma4, Llava, Phi4, ...) (AAAI 2025).

  • 1,708−11 Elo, change today
  • 16.1k stars
  • 1.8k forks
  • 55 watchers
git
git clone https://github.com/modelscope/ms-swift.git
v4.6.12d ago
  • deepseek
  • embedding
  • glm
  • grpo

LlamaChinese/Llama-Chinese

Open on GitHubPython

Llama中文社区，实时汇总最新Llama学习资料，构建最好的中文Llama大模型开源生态，完全开源可商用

  • 1,485−8 Elo, change today
  • 14.7k stars
  • 1.2k forks
git
git clone https://github.com/LlamaChinese/Llama-Chinese.git
  • agent
  • llama
  • llama4
  • llm

README

Loading README…